"However, when similar flight paths are compared across conditions, the stability of the hippocampal code persists."
Doesn't this mean that "place cell" is a misnomer (for such firing patterns)? They're really "place+trajectory" cells. Shouldn't a "true" place cell be path-invariant?
Credit-assigning on such invariant reps is presumably needed somewhere in the system in order to drive flexible navigation, no?
@DrYohanJohn Eichenbaum conceived of place cells more as spatial relation cells: distance relation to this or that wall, for example. I see they cite one of his papers (oldish one though), but I can't access the paper to see what they cite him as saying or the details of the differing conditions in question.
Right but distance is already always a relation, either between landmarks or from a point to some "origin", which is also basically a landmark. There's no other sense of "place", really.
And a pure "distance from X" cell would have trajectory-invariance.
@DrYohanJohn@WorldImagining
Eichenbaum's Law* states, "if you do an experiment to determine if neurons in the hippocampus code for task relevant variable X or Y, answer is always yes."
There are neurons tiling distance and neurons tiling time and neurons coding distance x time. (Same for many other variables ). If you want to think of the whole data structure you're forced to think of a pretty complex object. We talk about "place cells" or "time cells" (or or) because it's difficult to think about that object (and neuro people hate math). I don't think anyone seriously thinks "place cells" are a real thing separate from other variables (Howard certainly didn't), it's just a name that makes it easier to talk about experiments.
As far as I recall Howard never stated Eichenbaum's Law but he also did not object when I stated it and attributed it to him so I think it's ok .
This make sense to me, but I also think that from a normative perspective it would be nice if there were path-invariant place cells (or a distributed with this property), even if they were strongly context variant... under the assumption that credit-assignment for place preference/avoidance should be at least somewhat path-invariant. Such representations could then guide navigation with that context, given contingent constraints.
It also strikes me that very distributed representations might create problems for credit assignment: reinforcement would lead to a lot of feature-based overgeneralization.
In other words, it makes sense that conjunctive representations (logical AND like) exist and are relatively localized.
How might one design reinforcement on distributed reps that avoid/minimize such overgeneralization?
There are path-invariant place cells for positional credit assignment, time cells for temporal credit assignment etc etc. But i agree that this approach seems like it would be fraught with challenges and in any event doesn't really draw on the power of this kind of representation.
Maybe feature-based isn't the way to go. Maybe we shouldn't think about the individual cells and instead think about the relationships that are coded across the whole population by this rich, continuous, multidimensional data structure.
This paper is an attempt to approach this problem thinking about temporal credit assignment. In principle one could generalize to other continuous dimensions.
Add comment